AI as a Therapist: What It Really Says About Us at Work

When people use AI as a therapist, it is rarely a story about technology. It is a story about how unsafe we have made it to be human with each other. When a machine feels safer than a person, that is not a technology problem. It is a loneliness crisis, and it is quietly showing up at work.

More and more, people are opening up to a chatbot before they open up to a colleague, a manager, or a friend. They type the questions they cannot say out loud:

“I feel depressed but I don’t know why. Is that normal?”
“How do I stop feeling like I’m failing at everything?”
“Why do I feel anxious even when nothing is wrong?”

The real question isn’t whether the technology is clever. It’s this: why are people finding it easier to share their feelings with a machine than with another human being?

Why do people use AI as a therapist?

Because the machine does not interrupt. It does not give unwanted advice. It does not make it about itself. It doesn’t say, “you’re overthinking this.” It acknowledges what you feel. It tells you your feelings make sense. It offers a perspective. And then it allows you the freedom to take it or leave it. No judgement, just a safe space. Which is more than most humans are offering each other right now.

When a machine feels safer than a colleague

You can sit beside a colleague for months, and yet they know nothing about what you feel, or what you are struggling with. That gap is the real story. It isn’t that AI has become so advanced; it’s that our human spaces have become so guarded. This is where psychological safety matters most: when people don’t feel safe being honest, the honesty doesn’t disappear. It just moves to a screen. What looks like a technology trend is often loneliness wearing a modern disguise.

What AI as a therapist can, and can’t, do for us

Some of this is genuinely good. Being able to name a feeling, feel heard, and get a calmer perspective at 2am has real value, and there is no shame in reaching for it. But a chatbot can only take us so far. It can simulate being heard. It cannot actually know you, sit with you, or notice when you’ve gone quiet for a week. The deeper fix isn’t a better chatbot. It’s building workplaces where meaningful connection and humanising the workplace make it safe to be a person again.

If you, or someone on your team, is genuinely struggling, a chatbot is a starting point, not a safety net. Real support, from a trusted person or a qualified professional, still matters most.

Frequently asked questions